Thursday, 2024-04-04

gokhaniHello folks, I have created a non-ssl environment on public endpoint. I am getting certificates ssl very failed errors when trying to create a loadbalancer. how can we reslove this issue ? it is working on environments with ssl. 07:17
noonedeadpunkgokhani: loadbalancer in terms of Octavia?07:59
noonedeadpunkalso what's the error?08:00
noonedeadpunkeventually, as you might know, Octavia does still require to have TLS as authentication with Amphora happens through mTLS auth08:00
gokhaninoonedeadpunk: yes ı mean octavia, octavia amphora driver could not connect to instance 08:29
gokhanithere is self certificate but it is not working 08:29
noonedeadpunkgokhani: could it be that it got accidentally rotated ?08:31
gokhaninoonedeadpunk: May be after upgrade it is rotated. I am checking now 08:33
noonedeadpunkeventually, failovering loadbalancers should help if that is the case08:47
noonedeadpunkas client certificates are passed as metadata to amphora during spawn-up08:48
noonedeadpunkso once they're there - changing server certificate will lead to that situation08:48
noonedeadpunkloadbalancer failover does re-create VMs, so new certificate pair will be pulled08:49
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Switch service repos to track 2024.1  https://review.opendev.org/c/openstack/openstack-ansible/+/91418809:05
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Switch service repos to track 2024.1  https://review.opendev.org/c/openstack/openstack-ansible/+/91418809:06
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Add trove tempest testing  https://review.opendev.org/c/openstack/openstack-ansible/+/78437909:08
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-os_trove master: DNM  https://review.opendev.org/c/openstack/openstack-ansible-os_trove/+/91505209:09
gokhaninoonedeadpunk: I can not reach amphora instances with ssh. it requests password for login. 09:30
gokhanimay be it can not reach metadata service 09:31
noonedeadpunkgokhani: so there should be an SSH key on the deploy host ideally09:33
noonedeadpunkin case `octavia_ssh_enabled` is set to True09:34
noonedeadpunkhttps://opendev.org/openstack/openstack-ansible-os_octavia/src/branch/master/defaults/main.yml#L312-L31409:34
gokhaninoonedeadpunk: yes it is true. I am trying to ssh with this keypair. 09:38
gokhanibut instance requests password 09:40
gokhaniI think user is amphora or ubuntu 09:40
noonedeadpunkubuntu09:57
noonedeadpunkgokhani: does newly spawned LBs are also having the same issue?10:01
noonedeadpunkAnd have you tried just to failover LB?10:01
gokhaninoonedeadpunk: I couldn't failover because it is in pending create state 12:09
gokhanihow can we refresh octavia certs 12:09
noonedeadpunkgokhani: and can you failover just specific amphoras?12:19
gokhaninoonedeadpunk: I solved issue, after remove octavia migration user variable file and running octavia install it is resolved 12:36
noonedeadpunkhuh12:53
noonedeadpunkok12:53
gokhaninoonedeadpunk: without deleting them it skips install server certificates12:55
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Filter out empty src/source from user-role/collection-requirements  https://review.opendev.org/c/openstack/openstack-ansible/+/91507414:45
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: [doc] Document usage of user.rc file  https://review.opendev.org/c/openstack/openstack-ansible/+/91507615:25
opendevreviewMerged openstack/openstack-ansible-os_designate stable/2023.2: Fix designate upgrades when internal RPC version changes  https://review.opendev.org/c/openstack/openstack-ansible-os_designate/+/91487515:56
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: [doc] Rename extending-osa page  https://review.opendev.org/c/openstack/openstack-ansible/+/91507815:56
noonedeadpunkI will propose new 2023.2 bump as we've backported quite some new things lately...15:59
noonedeadpunkand I haven't proposed new minor releases yet15:59
noonedeadpunkwanna wait for this though: https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/91488516:00
noonedeadpunkjrosser_: capi fails weirdly now - https://zuul.opendev.org/t/openstack/build/a83cfacdfd664bc88c5b02ce0c3c101f/log/job-output.txt#2747916:06
noonedeadpunk| status               | CREATE_IN_PROGRESS     16:06
noonedeadpunk"| status_reason        | CAPI Cluster status: Provisioned: Cluster kube-2mplm is Provisioned. CAPI OpenstackCluster status reason: Successfulcreatemonitor: Created monitor k8s-clusterapi-cluster-magnum-system-kube-2mplm-kubeapi-6443 with id 9bfa0814-f0e6-414f-8ef6-c95a0f7d42c0 |16:06
noonedeadpunkwhich doesn't really add-up for me16:06
jrosser_well - getting the cluster provisioning status to be transferred over to the status in magnum is a thing that has to go right16:13
noonedeadpunkjust reason looks like it should be completed I assume?16:14
noonedeadpunkor maybe not fully16:15
opendevreviewDmitriy Rabotyagov proposed openstack/ansible-role-uwsgi master: Add Debian 12 distro setup variable  https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/91508016:22
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add distro infra jobs  https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/91469116:22
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Implement installation method selection for MariaDB role  https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/91453016:23
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add distro infra jobs  https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/91469116:23
jrosser_it’s going to be a week before I can take a look at that16:23
opendevreviewMerged openstack/openstack-ansible-plugins master: Add check_hostname option to db_setup tasks  https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/90021716:44
spatelnoonedeadpunk quick question, in horizon AVAILABLE_REGIONS option I can specify multiple region but how does auth will handle here if I have totally isolated clouds? 16:45
spatelI have two individual cloud and I would like to manage them with single horizon UI 16:46
noonedeadpunkthere was another variable there16:46
spatel?16:46
noonedeadpunkhttps://opendev.org/openstack/openstack-ansible-os_horizon/src/branch/master/defaults/main.yml#L262-L26416:46
noonedeadpunkBasically you can supply pair of keystone/region name16:46
spatelYes but how does auth works? if I select different region then how there region auth or pass token? 16:47
noonedeadpunkso AVAILABLE_REGIONS is actually list of mappings, where you provide keystone url and region name as second key16:47
noonedeadpunkit's not?:)16:47
noonedeadpunkor well, you need to auth towards proper regions/keystone16:48
noonedeadpunkif you have a valid token in cookies - probably you can switch back and forth16:48
noonedeadpunk(for each region)16:48
noonedeadpunkso you auth towards each region independently16:49
noonedeadpunkunless you've stretched keystone :)16:49
spatelI don't have stretched keystone :(16:50
spatelboth cloud are isolated 16:50
spatelThat is why I am asking question how does multi-region will work here if both doesn't know about each other 16:50
noonedeadpunkI mean, it's not really multi-region:)16:51
noonedeadpunkso you'd need to auth independently when switching regions16:52
spatelso we have to type password :)16:52
spatelyike.. 16:52
noonedeadpunkor share keystone. or do federation16:54
noonedeadpunks/share/stretch/16:54
spatelis federation easy to manage? 16:54
noonedeadpunkUm. Might be. I don't have positive experience yet though16:55
spatelor how about dump keystone table from A to B region using ansible :) 16:55
noonedeadpunkI think I wrote about that lately, but scheme with strwetching a separate galera cluster cross-region for keystone - sounds appealing enough to me16:56
noonedeadpunkgoing to play with that in practice soonish16:56
spateljust sync keystone DB stuff? 16:56
spatelI am thinking what if just dump user/pass/role etc instead entire DC16:57
noonedeadpunkyeah. and well - you'd need to sync also fernets through SSH in case of osa. But that's minor16:57
noonedeadpunkTokens depend on fernets16:57
spatelits easy to do right?16:57
noonedeadpunkSo if fernets are not in sync - tokens are not valid16:57
opendevreviewMerged openstack/openstack-ansible-os_designate stable/2023.1: Fix designate upgrades when internal RPC version changes  https://review.opendev.org/c/openstack/openstack-ansible-os_designate/+/91487618:13
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-plugins stable/2023.2: Do not log contents of installed keypairs by default  https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/91501620:09
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-plugins stable/2023.1: Do not log contents of installed keypairs by default  https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/91501720:09
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone master: Fix permissions for SSH private key for Ubuntu distro installations  https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/91508920:38
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone master: Fix permissions for SSH private key for Ubuntu distro installations  https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/91508920:38
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add distro infra jobs  https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/91469120:39

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!